Echoic and retrieval accounts of the long-term modality effect tested using the suffix procedure

The long-term modality effect is the advantage in recall of the last of a list of auditory to-be-remembered (TBR) items compared with the last of a list of visual TBR items when the list is followed by a filled retention interval. If the auditory advantage is due to echoic sensory memory mechanisms, then recall of the last auditory TBR item should be substantially reduced when it is followed by a redundant, not-to-be-recalled auditory suffix. Contrary to this prediction, Experiment 1 demonstrated that a redundant auditory suffix does not significantly reduce recall of the last auditory TBR item. In Experiment 2 a nonredundant auditory suffix produced a large reduction in the last auditory item. Redundancy is not the only factor controlling the effectiveness of a suffix, however. Experiment 3 demonstrated that a nonredundant visual suffix does not reduce recall of the last auditory TBR item. These results are discussed in reference to a retrieval account of the long-term modality effect.     

A test of precategorical and attentional explanations of speech suffix interference

It is well-known that a redundant item or “stimulus suffix”, which does not have to be recalled, which is spoken, and which terminates the presentation of a beyond span-length sequence of to-be-remembered (TBR) spoken items, will generate considerable interference in immediate serial recall. Previous work has established that speech suffix interference is not influenced by a number of intrinsic attributes of the suffix. A limitation of this work, however, is the fact that in each study there has been little if any disparity between TBR sequence and suffix in terms of extrinsic attributes. Experiment I extends this work and showed that at least one intrinsic attribute of the suffix (personal significance) has no effect on suffix interference when there is also a marked disparity between TBR sequence and suffix in terms of one specific extrinsic attribute (spatial location). Experiment II simply served to eliminate one possible artifactual explanation of this finding. The result appears to have some theoretical importance since it is inconsistent with Kahneman's (1973) attentional account of suffix interference and with one interpretation of a precategorical hypothesis due to Morton, Crowder and Prussin (1971).     

The effects of rime on auditory recency and the suffix effect

Two experiments are reported that consider the role of rime as the content of Precategorical Acoustic Storage (PAS). It was hypothesised that with auditory presentation of lists the rime component of the final item (the final vowel and, optionally, terminal consonant cluster of a word) was preserved in PAS and this served as a recall cue to identify the final item. Experiment 1 compared recall performance when the terminal or penultimate words rhymed with an irrelevant stimulus suffix. A suffix that shared a rime with the terminal word was found to attenuate the suffix effect, in line with Carr and Miles (1997) and consistent with the use of PAS as a positional code. However, contrary to the hypothesis that PAS information is automatically used to reconstruct the final item, the suffix effect was no greater when the suffix rhymed with the penultimate item than when no relationship existed between the suffix and the final item. Experiment 2 demonstrated that when the terminal and penultimate words shared a rime there was a drop in recall performance for the final item but a corresponding increase in correct recall of the penultimate item. No such changes were observed when the rime was shared between terminal and antepenultimate items. It is suggested that adjacency of identical rimes allows improved reconstruction of item information at the expense of order information.     

Two-component theory of the suffix effect: Contrary evidence

The suffix effect is the selective impairment in recall of the final items of a spoken list when the list is followed by a nominally irrelevant speech item, or suffix. It is widely assumed to comprise a bottom-up, or structural, effect restricted to the terminal item and a top-down, or conceptually sensitive, effect confined to the preterminal items. Reported here are eight experiments that challenge this view by demonstrating that the terminal suffix effect, as well as the preterminal suffix effect, is susceptible to conceptual influence. The entire suffix effect may be better conceived of as a phenomenon arising from perceptual grouping.     

Similarities between the irrelevant sound effect and the suffix effect

Although articulatory suppression abolishes the effect of irrelevant sound (ISE) on serial recall when sequences are presented visually, the effect persists with auditory presentation of list items. Two experiments were designed to test the claim that, when articulation is suppressed, the effect of irrelevant sound on the retention of auditory lists resembles a suffix effect. A suffix is a spoken word that immediately follows the final item in a list. Even though participants are told to ignore it, the suffix impairs serial recall of auditory lists. In Experiment 1, the irrelevant sound consisted of instrumental music. The music generated a significant ISE that was abolished by articulatory suppression. It therefore appears that, when articulation is suppressed, irrelevant sound must contain speech for it to have any effect on recall. This is consistent with what is known about the suffix effect. In Experiment 2, the effect of irrelevant sound under articulatory suppression was greater when the irrelevant sound was spoken by the same voice that presented the list items. This outcome is again consistent with the known characteristics of the suffix effect. It therefore appears that, when rehearsal is suppressed, irrelevant sound disrupts the acoustic-perceptual encoding of auditorily presented list items. There is no evidence that the persistence of the ISE under suppression is a result of interference to the representation of list items in a postcategorical phonological store.     

Serial recall of two-voice lists: Implications for theories of auditory recency and suffix effects

Substantial recency effects are found in immediate serial recall of auditory items. These recency effects are greatly reduced when an irrelevant auditory stimulus (a stimulus suffix) is presented. A number of accounts that have been proposed to explain these phenomena assume that auditory items are susceptible to masking or overwriting in memory. Later items overwrite earlier items, leading to an advantage for the last item, unless it is masked by a suffix. This assumption is called into question by evidence that presenting list items in two voices has no beneficial effect in immediate serial recall. In addition, it is shown that suffix effects on both terminal and preterminal list items are influenced by the physical similarity of the suffix to the terminal item and not by the physical similarity of the suffix to preterminal items.     

The role of memory in attenuations of the suffix effect

Various properties attenuate the auditory suffix effect. These properties may improve recall by allowing the suffix to be excluded from the auditory store (or stores) used for immediate ordered recall, or they may improve recall even though the suffix is included in the auditory store. Nine items were presented, with the last item being either a suffix or another item to be recalled. In Experiment 1, the impairment in recall of the previous items produced by presenting a to-be-recalled last item in a different voice was the same as the impairment produced by presenting a suffix in a different voice. The same result was found in Experiment 2 for presenting the last item in a different spatial location and in Experiment 3 for delaying the presentation of the last item. Apparently the attenuating properties improve recall through memorial factors, and not by allowing the suffix to be excluded from the auditory store.     

Echoic memory and voice quality: Recency recall is not enhanced by varying presentation voice

Three experiments explored the effects on immediate recall of varying voice of presentation. Experiment 1 showed that the free-recall recency effect was not enhanced by presenting list words alternately in a male and a female voice. Experiment 2 replicated this result and also showed that recall of a given recency item from such a list was no more probable when the subjects were informed immediately following presentation that they need not recall the words presented in the other voice. Experiment 3 replicated previous findings of a reduction in the “suffix effect” when presentation voice is changed for the suffix item. The relation of this result to those of Experiments 1 and 2 is discussed.     

Two-component theory of the suffix effect: contrary findings

The suffix effect is the reduction in the recallability of the last few items of a just-spoken list caused by appending a nominally irrelevant item, or suffix. The effect is widely assumed to comprise a "structural" terminal component, affecting just the last item, and a strategy-sensitive preterminal component. In a series of 8 experiments, the authors fail to replicate any of a variety of findings widely cited as the empirical basis for this 2-component theory. The authors also question the support that earlier findings provide for the theory, even if they had proved replicable. They attribute the entire suffix effect to the grouping of the suffix with the list items.     

Recency and suffix effects in serial recall of musical stimuli

Auditory presentation of verbal items leads to larger recency effects in recall than visual presentation. This enhanced recency can be eliminated if a stimulus suffix (an irrelevant sound) follows the last item. Four experiments tested the hypothesis that recency and suffix effects in serial recall result from a speech-specific process. It was demonstrated that serial recall of musical notes played on a piano exhibited substantial recency effects. These recency effects were reduced when the list items were followed by either a piano chord or the word start. However, a white-noise suffix had no effect on recency. This pattern of data is consistent with current work on auditory perception and places constraints on theories of recency and suffix effects.     

Adapting to an irrelevant item in an immediate recall task

Recall of the last one or two items of a spoken list is impaired when the list is followed by a nominally irrelevant item. At issue here was whether this suffix effect is reduced with repeated exposure to the irrelevant item. The effect was found to decline over successive blocks of trials, but only slightly (Experiment 1). No decisive evidence for adaptation to the irrelevant item was found when it was spoken after each of the list items rather than after the last one only (Experiments 2 and 3). The strongest evidence for adaptation was obtained when the irrelevant item was repeated in an unbroken stream that extended through the presentations and recall periods of successive lists: The recency effect and the level of recall at the last position within a list were greater under these conditions than when the irrelevant item was presented only once after each list (Experiments 4, 5, and 6).     

Maintaining recency despite a stimulus suffix

Two processes of information retrieval were considered in the context of the logogen model. The aim was to establish whether information about the final items of an auditory short-term memory list is held exclusively in precategorical acoustic storage at presentation or whether these items are automatically registered in a cognitive store as well. Error data for a final heterogeneous item in alphanumeric lists showed significantly better recall, despite the addition of a stimulus suffix. Although these results demonstrated that coding had proceeded further than a precategorical stage, which maintains only physical features, the possibility remained that the effect was due to a bias in focal attention and selective coding at list presentation. A second experiment increased the difficulty of the retrieval task, and effectively precluded the possibility of a bias in attention. The results confirmed the findings in the first experiment. It was concluded that information about the final item(s) is registered automatically in the cognitive system, and that responses are made available from this source when information about physical features of the item is degraded by a stimulus suffix.     

Suffix effects and probe modality in probed recall: Implications for readout from sensory memory

The effect of a redundant stimulus suffix was investigated in three experiments using a probed recall task. The probe modality was auditory in the first experiment and visual in the second. In the third experiment, both probe modalities were tested. The presentation rate was varied in all three experiments and was confounded with the delay of the suffix. The suffix was found to have a small and approximately equal effect on the last six serial positions. In contrast to the predictions from Crowder and Morton's (1969) PAS model, presentation rate and suffix delay did not interact with the suffix effect. The results indicate that at least six items can be simultaneously represented in PAS, that there is no appreciable decay of information from PAS during presentation of the last list items, and that readout of PAS information does not occur either during list presentation or between the end of list presentation and the beginning of overt report.     

The delayed stimulus suffix effect following arhythmic stimulus presentation

In auditory memory span experiments an extra word presented at the end of the list impairs performance on the last few items even though it is completely redundant; delaying this stimulus suffix past the time when an additional memory item would have occurred--so that the rate of presentation is retarded for the suffix--reduces the magnitude of the performance impairment. One account of this pattern of results is that during presentation of the items the subject adjusts periodic bursts of attention to the regular cadence of new information. Since an immediate suffix is in time with this cadence and a delayed suffix not, the larger effect of the former is explained without appeal to any process dependent on real time. In the present study the stimulus items were presented in an arhythmic manner with interstimulus intervals ranging from 100 to 900 ms on a random basis. Under these conditions the same effect of delaying the suffix was observed as is typical of rhythmic stimulus presentation; therefore, the rhythm hypothesis for the effect is not supported.     

Short-Term Visual Recognition Memory in Pigeons

Recognition memory for lists of items was investigated in pigeons using a YES-NO recognition technique. Experiment I showed that increasing the exposure duration of the first item of a two-item list improved recognition for that item without impairing recognition of the second item. Experiment II showed that decreasing the inter-trial interval had no effect on correct YES responses but significantly increased the number of false YES responses. Experiment III showed that recognition for the last two items of a three-item list was no poorer than that for lists of only two items. Experiment IV showed that increasing the delay between presentation and test of a two-item list (from 0·25-1 s) had a more disruptive effect on recognition for the second than for the first item. The data from these four experiments support a model proposed by Roberts and Grant, according to which memory traces are independent, and decay as a negatively accelerated function of time. Experiments V, VI, and VII investigated recognition for lists of three, four, and five items, and found no evidence for a primacy effect, performance being a linear function of time since sample offset.     

Differential effects of repetition on pre- and postcategorical memory traces

The effect of list repetition on immediate recall for aurally presented nine-letter lists was studied under two conditions. In the first, a redundant stimulus item was presented as the tenth item in each list, while the suffix was not included in a control condition. As in previous research (Crowder & Morton, 1969), the stimulus suffix selectively interfered with recall at the terminal presentation serial positions, indicating the presence of precategorical acoustic storage. Repetition had a nonselective effect on performance. This result and an analysis of acoustic errors support the inference that qualitative differences in the memory code may lead to differences in other functional properties of the memory trace, such as responsiveness to repetition.     

Capturing the suffix: cognitive streaming in immediate serial recall

Adding an irrelevant item to the end of an auditory to-be-remembered list increases error on the last list items appreciably, known as the suffix effect. The phenomenon of auditory capture (e.g., Bregman & Rudnicky, 1975), namely, the tendency for a sequence of similar items to form a stream that at the same time isolates perceptually dissimilar members of the sequence, is exploited to explore the suffix effect. Irrelevant items interleaved between to-be-remembered items are used to capture the suffix with the aim of reducing its impact. Four experiments illustrate how the properties of the irrelevant sequence promote capture. The results are problematic for models of the suffix that involve masking of the last list item; instead, models based on grouping are favored.     

SHORT REPORT Rhyme Attenuates the Auditory Suffix Effect: Alliteration Does Not

Two experiments are reported in each of which subjects were required to recall spoken consonant-vowel-consonant (CVC) syllable lists serially. Both experiments contrasted the effects of a rhyming suffix and an alliterative suffix on recall of the terminal list item. A rhyming suffix reliably attenuated the normally robust suffix effect; an alliterative suffix did not. The finding points to the importance of the location rather than the quantity of phonological repetition in determining the size of the suffix effect. In line with Treiman and Danis (1988), it is argued that the onset (C) and rime (VC) components of CVC syllables may exist as separate entites within short-term acoustic memory. This, coupled with the superior durability of the rime component within acoustic memory, affords the subject a greater probability of recalling correctly the terminal list item in the rhyming suffix condition.     

Absence of a cross-modal “suffix effect” in short-term memory

Three experiments are reported involving the presentation of lists of either letters or digits for immediate serial recall. The main variable was the presence or absence of a suffix-prefix, an item (tick or cross) occurring at the end of the list which had to be copied before recall of the stimulus list. With auditory stimuli and an auditory suffix-prefix there was a large and selective increase in the number of errors on the last few serial positions—the typical “suffix effect”. The suffix effect was not found with auditory stimuli and a visual suffix-prefix nor with a visual stimulus and an auditory suffix-prefix. These results are interpreted as supporting a model for short-term memory proposed by Crowder and Morton (1969) in which it is suggested that with serial recall information concerning the final items following auditory presentation has a different, precategorical, origin from that concerning other items.     

Manipulations of irrelevant information: Suffix effects with articulatory suppression and irrelevant speech

Numerous studies have demonstrated impaired recall when the to-be-remembered information is accompanied or followed by irrelevant information. However, no current theory of immediate memory explains all three common methods of manipulating irrelevant information: requiring concurrent articulation, presenting irrelevant speech, and adding a stimulus suffix. Five experiments combined these manipulations to determine how they interact and which theoretical framework most accurately and completely accounts for the data. In Experiments 1 and 2, a list of auditory items was followed by an irrelevant speech sound (the suffix) while subjects engaged in articulatory suppression. Although articulatory suppression reduced overall recall compared to a control condition, comparable suffix effects were seen in both conditions. Experiments 3 and 4 found reliable suffix effects when list presentation was accompanied by irrelevant speech. Experiment 5 found a suffix effect even when the irrelevant speech was composed of a set of different items. Implications for working memory, pre-categorical acoustic store, the changing-state hypothesis, and the feature model are discussed.